The present invention relates generally to network management, and more particularly, to configuring the contents of an SNMP notification.
It is desirable to manage and control nodes of a network such as a TCP/IP network from one or more network management stations connected to the network. The network management stations monitor and control other nodes in the network including hosts, routers, etc. A protocol known as the Simple Network Management Protocol (SNMP) is used to communicate management information between network management stations (NMS) and SNMP agent software operating on other network nodes, or the same node as the network management station. SNMP is described in Case, RFC 1157, “A Simple Network Management Protocol (SNMP),” (Internet Engineering Task Force May 1990), the contents of which are incorporated by reference herein. Using SNMP in a TCP/IP network, a network management station may monitor traffic at a remote node and modify network operation parameters at the remote node.
FIG. 1 is a block diagram illustrating an exemplary network utilizing SNMP. The network includes at least one management station 10 which controls a plurality of nodes such as workstation 11, bridge 12, router 14, printer 15, or other peripheral devices (e.g., personal computers, hubs, repeaters). The management station 10 includes a SNMP manager 16 (command generator and notification receiver). In order for an SNMP management station 10 to manage a node, the node must be able to run an SNMP agent 17 (command responder and notification originator), which is an SNMP management process typically encoded in software (some nodes may use a hardware implementation).
To facilitate SNMP operation, nodes of the network including the network management stations 10 maintain network management information databases known as MIBs (management information bases) 18. MIBs are described in McCloghrie, et al., RFC 1213, “Management Information Base for Network Management of TCP/IP Based Internets: MIB-II” (Internet Engineering Task Force March 1991), the contents of which are incorporated by reference herein.
The MIB 18 at each node consists of many individual objects, each having one or more instance. Each instance has a single value. For example, a particular node may include within its MIB an object whose value specifies the total number of IP datagrams received at that node, including those received in error. Each MIB 18 includes a large number of such objects. The object instances and their corresponding values are carried in packets called protocol data units (PDUs) and contain operating parameters, statistics, and control information for the element and its components.
Each MIB object is uniquely identified by a series of integers called an Object Identifier (OID). For example, the object ifinOctets is represented by the OID 1.3.6.1.2.2.1.10. The can be many instances of an object in existence on an agent. For example, there is one instance (and value) of ifInOctects for every physical interface known to the agent. Each instance is identified by appending one or more additional integers to the OID (e.g., 1.3.6.1.2.1.2.2.1.10.x, where x is an integer greater than zero). The first element of this specific OID identifies an overall object identifier domain allocated to the ISO. The second element has the value 3 which is defined as indicating allocation toward an organization under the auspices of the ISO. The third element identifies the US Department of Defense (DOD) as the responsible organization. The fourth element has the value of 1 and identifies the Internet. The fifth element indicates that the identifier is used for management. The remaining elements identify the particular object types with greater specificity.
The organization of objects within an MIB may be illustrated as a tree structure. An example of a portion of such a tree structure is illustrated in FIG. 2. The leaves of the tree instances of individual objects. FIG. 2 shows the MIB tree from its root, “ISO”, to some of its lower branches. A first branch 21 of primary interest includes the standard MIB objects defined by RFC 1213. A second branch 23 includes MIB objects defined for use by Cisco Systems. The illustration of FIG. 2 does not extend down to individual leaves but rather indicates the prefix of elements in OIDs for various classes of objects. For example, OIDs from MIB objects pertaining to TCP would begin with 1.3.6.1.2.1.6.
The management station communicates with the agents over the network using the SNMP protocol, which allows the management station to query the state of the agent's local objects and modify them if necessary. SNMP is a request-response protocol by which the variables of an agent's MIB may be inspected or altered. The protocol is described in RFC 1905, “Protocol Operations for Version 2 of the Simple Network Management Protocol”, Case, McCloghrie, Rose & Waldbusser, January 1996, the contents of which are incorporated by reference herein.
A typical SNMP operation involves management station sending an SNMP message to agent requesting a particular local parameter. The agent then recovers this parameter by accessing a particular object in the MIB and responds to the request with a reply to the management station including the parameter value. SNMP operations are defined on a list of MIB variable instances called a variable binding (varbind) list. Each element of a varbind list describes a specific MIB variable instance. A varbind element specifies three MIB instance attributes: its object identifier, data type, and value.
The management protocol provides for the exchange of messages which convey management information between the agents and management stations. For example, the management station may send a request to an agent asking it for information or commanding it to update its state in a certain way. The agent typically replies with the requested information or confirms that it has updated its state as requested. Conventional message types include Get, GetNext, GetBulk, Set, Trap, and Inform. The Get operation is used by the system to retrieve the value of one or more object instances from an agent. The GetNext operation is used by the system to retrieve the value of the next object instance in a table or list within an agent. The GetBulk operation is used by the system to efficiently retrieve large blocks of data, such as large tables. The Set operation is used by the network management station to set the values of object instances within an agent. The Trap operation is used by agents to asynchronously inform the network management station of a event of interest to one or more network management stations. Notification originators can generate informs, which are notifications that should be acknowledged by notification receivers. In an acknowledgement is not received within a configurable time-window, the notification originator attempts to resend the inform up to a maximum of N returns, where N is a configurable integer. The Trap and Inform operations may be used to send notifications to the manager.
Notifications are unsolicited messages sent from an agent to the SNMP manager to apprise the manager of network events. SNMP notifications allow the agent to initiate communication with management applications when an event of interest to one or more network management stations takes place. These events include cold start, warm start, link down, link up, authentication failure, neighbor loss, and enterprise. SNMP notifications contain a sequence of SNMP objects that typically provide context information on why each notification was generated.
Many networks contain components manufactured by several different companies. In order for the management station to communicate effectively with these varied devices, the information maintained by the agents must be rigidly specified. SNMP therefore describes the exact information each agent 17 must maintain and the format in which it must be maintained in the MIB 18. At times, newer versions of MIBs result in new objects being added to notifications. Also, notifications are often generated with different variable bindings than those that should be listed in the MIB, either in error or because a network element has not kept up with newer versions of MIBs. Furthermore, it is often desirable to add specific objects to a notification to provide additional flexibility in management of a network.
There is, therefore, a need for a system and method for instructing an agent which objects to send for each notification so that notifications can be easily reconfigured by a management station.